boofcv.abst.feature.detect.line

Class DetectLineHoughFootSubimage<I extends ImageSingleBand,D extends ImageSingleBand>

java.lang.Object

boofcv.abst.feature.detect.line.DetectLineHoughFootSubimage<I,D>

All Implemented Interfaces:

DetectLine<I>

public class DetectLineHoughFootSubimage<I extends ImageSingleBand,D extends ImageSingleBand>
extends Object
implements DetectLine<I>

Detects lines inside the image by breaking it up into subimages for improved precision. Inside each subimage a hough transform is independently computed. See [1] for more details.

USAGE NOTES: Blurring the image prior to processing can often improve performance. Results will not be perfect and to detect all the obvious lines in the image several false positives might be returned.

[1] Section 9.3 of E.R. Davies, "Machine Vision Theory Algorithms Practicalities," 3rd Ed. 2005

Author:

Peter Abeles

See Also:

HoughTransformLineFootOfNorm

Constructor Summary

Constructors

Constructor and Description
DetectLineHoughFootSubimage(int localMaxRadius, int minCounts, int minDistanceFromOrigin, float thresholdEdge, int totalHorizontalDivisions, int totalVerticalDivisions, int maxLines, ImageGradient<I,D> gradient) Specifies detection parameters.

Method Summary

Methods

Modifier and Type	Method and Description
List<georegression.struct.line.LineParametric2D_F32>	detect(I input) Detect lines inside the image.
ImageUInt8	getBinary()
D	getDerivX()
D	getDerivY()
ImageFloat32	getEdgeIntensity()
HoughTransformLineFootOfNorm	getTransform()

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

DetectLineHoughFootSubimage

public DetectLineHoughFootSubimage(int localMaxRadius,
                           int minCounts,
                           int minDistanceFromOrigin,
                           float thresholdEdge,
                           int totalHorizontalDivisions,
                           int totalVerticalDivisions,
                           int maxLines,
                           ImageGradient<I,D> gradient)

Specifies detection parameters. The suggested parameters should be used as a starting point and will likely need to be tuned significantly for each different scene.

Parameters:

localMaxRadius - Lines in transform space must be a local max in a region with this radius. Try 5;

minCounts - Minimum number of counts/votes inside the transformed image. Try 5.

minDistanceFromOrigin - Lines which are this close to the origin of the transformed image are ignored. Try 5.

thresholdEdge - Threshold for classifying pixels as edge or not. Try 30.

maxLines - Maximum number of lines it will detect. Try 10.

gradient - Computes the image gradient.

Method Detail

detect

public List<georegression.struct.line.LineParametric2D_F32> detect(I input)

Description copied from interface: DetectLine

Detect lines inside the image.

Specified by:

detect in interface DetectLine<I extends ImageSingleBand>

Parameters:

input - Input image.

Returns:

List of found lines.

getTransform

public HoughTransformLineFootOfNorm getTransform()

getDerivX

public D getDerivX()

getDerivY

public D getDerivY()

getEdgeIntensity

public ImageFloat32 getEdgeIntensity()

getBinary

public ImageUInt8 getBinary()